TY - CHAP
T1 - Image registration for interventional MRI guided procedures
T2 - Interpolation methods, similarity measurements, and applications to the prostate
AU - Fei, Baowei
AU - Lee, Zhenghong
AU - Duerk, Jeffery L.
AU - Wilson, David L.
PY - 2003
Y1 - 2003
N2 - Nuclear medicine can detect and localize tumor in the prostate not reliably seen in MR. We are investigating methods to combine the advantages of SPECT with interventional MRI (iMRI) guided radiofrequency thermal ablation of the prostate. Our approach is to first register the low-resolution functional images with a high resolution MR volume. Then, by combining the high-resolution MR image with live-time iMRI acquisitions, we can, in turn, include the functional data and high-resolution anatomic information into the iMRI system for improved tumor targeting. In this study, we investigated registration methods for combining noisy, thick iMRI image slices with high-resolution MR volumes. We compared three similarity measures, i.e., normalized mutual information, mutual information, and correlation coefficient; and three interpolation methods, i.e., re-normalized sine, tri-linear, and nearest neighbor. Registration experiments showed that transverse slice images covering the prostate work best with a registration error of ≈ 0.5 mm as compared to our volume-to-volume registration that was previously shown to be quite accurate for these image pairs.
AB - Nuclear medicine can detect and localize tumor in the prostate not reliably seen in MR. We are investigating methods to combine the advantages of SPECT with interventional MRI (iMRI) guided radiofrequency thermal ablation of the prostate. Our approach is to first register the low-resolution functional images with a high resolution MR volume. Then, by combining the high-resolution MR image with live-time iMRI acquisitions, we can, in turn, include the functional data and high-resolution anatomic information into the iMRI system for improved tumor targeting. In this study, we investigated registration methods for combining noisy, thick iMRI image slices with high-resolution MR volumes. We compared three similarity measures, i.e., normalized mutual information, mutual information, and correlation coefficient; and three interpolation methods, i.e., re-normalized sine, tri-linear, and nearest neighbor. Registration experiments showed that transverse slice images covering the prostate work best with a registration error of ≈ 0.5 mm as compared to our volume-to-volume registration that was previously shown to be quite accurate for these image pairs.
UR - http://www.scopus.com/inward/record.url?scp=0142152384&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=0142152384&partnerID=8YFLogxK
U2 - 10.1007/978-3-540-39701-4_34
DO - 10.1007/978-3-540-39701-4_34
M3 - Chapter
AN - SCOPUS:0142152384
SN - 3540203435
SN - 9783540203438
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 321
EP - 329
BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
A2 - Gee, James C.
A2 - Maintz, J. B. Antoine
A2 - Vannier, Michael W.
PB - Springer Verlag
ER -